Aptos has introduced an Encrypted Mempool, an optional feature that hides transaction details until their order is finalized. The move targets frontrunning and sandwich attacks — the practice of bots spotting a pending trade and jumping ahead of it to profit from the price move it causes.
The feature doesn't force anything on developers or users. It's a tool they can switch on if they want protection from the kind of predatory behavior that's become routine on public blockchains.
How the Encrypted Mempool works
On most networks, transactions sit in a public waiting area called the mempool before they're confirmed. Anyone can see what's queued up — the asset, the size, the target price. That visibility is what makes sandwich attacks possible. A bot spots a large buy order, places its own buy just ahead of it, lets the original order push the price up, then sells into the spike. The trader who sent the original order gets a worse fill.
Aptos' Encrypted Mempool keeps those details out of sight until the transaction's position in the block is locked in. By the time anyone can see what's happening, the ordering is already settled. Bots can't insert themselves into a trade they can't see coming.
The design is optional, which matters. Some applications rely on transparent mempools for their own logic — arbitrage strategies, certain types of market making, or tools that simulate trades before committing. Those developers can leave the feature off. Others, particularly consumer-facing exchanges and wallets, may choose to enable it to offer better execution to their users.
A different approach to MEV
Frontrunning is a subset of what the industry calls MEV — maximal extractable value. It's the profit that validators, block builders, and searchers can squeeze out of a block by choosing what goes in and in what order. On Ethereum and other networks, MEV has turned into a multi-billion-dollar shadow industry, with relay networks and specialized auctions deciding whose transactions get included first.
Most attempts to tame it have come from outside the base layer — private transaction relays, commit-reveal schemes, or order-flow auctions that try to redistribute the profits. Aptos is taking a different route: baking the privacy into the chain's own transaction pipeline.
That doesn't eliminate MEV entirely. Validators still control ordering, and encrypted transactions can still be delayed or dropped. But it removes the simplest and most damaging form of it — the kind where a bot sees your trade and punishes you for it.
What developers get
For teams building on Aptos, the Encrypted Mempool is another lever to pull when configuring how their applications interact with the network. It's not a default setting, and it's not a cure-all. It won't stop validators from reordering transactions for their own benefit, and it doesn't address other forms of MEV like backrunning — where a bot places a trade immediately after yours to capture the price move you created.
Still, sandwich attacks are the ones retail users feel most directly. They show up as worse prices on swaps, unexplained slippage, and a vague sense that the deck is stacked. Hiding pending transaction data until ordering is final removes the window those bots depend on.
The feature is live now. Developers who want it can enable it in their application configuration. Those who don't can leave things as they are. Aptos hasn't said whether the feature will eventually become a default or whether other privacy enhancements are in the works.
What's clear is the trade-off: encrypted mempools add latency and complexity to the transaction flow, and they reduce the transparency that some on-chain tools rely on. Whether the protection is worth the cost is a decision each project will have to make on its own.
